Study of a melanic pigment of Proteus mirabilis

© INSTITUTPASTEUR/ELSEVIER Paris 1996

Res. Microbiol.

1996, 147, 167-t74

Study of a melanic pigment of Proteus mirabilis A. Agodi (~), S. Stefani t2), C. Corsaro o), F. Campanile (2), S. Gribaldo (o and G. Sichel o)(') ~I) Institz, te of General Biology and ~2~Chair of General Microbiology, University of Catania, via Androne 81, 95124 Catania (Italy)

SUMMAR~ The present study sought to determine whether the pigment produced by Proteus mirabilis front the L-forms of various aromatic amino acids under aerobic conditions is melanic in nature. It is a black-brown pigment which behaves like a melanin in many respects, namely solubility features, bleaching by oxidizing agents and positive response to the Fontarm-Masson assay. In the present study, for the first time, it was shown by electron spin resonance analysis that a bacterial melanin is aide to act as a free radical trap, as was previously demonstrated for other melanin.~ Scanning electron microscopy studies showed a specific organized structure of the pigment as rounded aggregates of spherical bodies. DNA hybridization data did not reveal, in the P. m/rab/I/s ganome, any nudeotide sequence related to Snewanella colwelliana me1,4, one of the two melanogenasls systems already defined at the molecular level in bacteria. Results obtained from exlpedments on pigment production inhibition suggest a possible role of twosinase in P. m/rab/#s melanogenesis. In conclusion, from the bulk of our Tssults, it appears that the pigment produced by P. mirabilis is melanic in nature.

Key-words: Pigment, Melanin, Proteus mirabilis, Melanogenesis, Tyrosine; Ultrastructure, Free radical traps, Tyrosinase.

INTRODUCTION Black-brown pigments with characteristics of melanin are produced by a wide variety of prokaryotic and eukaryotic organisms. Although melanin formation has been reported for numerous bacteria, it has been studied intensively in only a few species, including Vibrio tyrosinaticus (Pomerantz and Murthy, 1974), mutants of Vibrio cholerae (lvins and Holmes, 1981) and several species of Streptomyces (Larch and Ettlinger, 1972; Katz et al., 1983; Yoshimoto et aL, 1985).

Submitted June 18, 1995, accepted September 27, 1995. (*) Correspondingauthor.

Only two bacterial melanogenesis systems have been defined at the molecular level: the tyrosinase-encoding genes from Streptomyces sp. (Bernan et al., 1985; Huber et aL, 1985; Muller et al., 1988) and the melanin-biosynthesis-encoding operon (mel) from the marine bacterium Shewanella colweltiana (Fuqua et al., 1991). Bacteria of the Proteus-Providencia-MorganeUa group produce a brownish "melanin-like" pigment from the L-forms of various aromatic amino acids in the presence of iron salts under aerobic conditions (Muller, 1985): however, the pigment has not

168

A. AGODI ET AL.

tilled water for 24 h, and lyophilized. Samples were deproteinized by digestion with 37 % (v/v) HC1 for 7 days at room temperature and flocculated at 60°C for 1 h (Nicolaus et aL, 1964). The resulting pigment was washed to neutrality with distilled water and dried.

been properly characterized. The purpose of the present study was to perform the chemical and ultrastructural characterization of the pigment produced by Proteus mirabilb strain 297V isolated from a human source and to investigate its possible role and significance.

Pigment characterization The purified pigment from P. mirabilis 297V cultures was subjected to elemental analysis by standard methods (Analytisehe Laboratorien, Gummersbach, Germany). A sample of liver melanin isolated from Rana esculenta L, by the same procedure was included as a control. Solubility of the isolated pigment was tested with alkaline water (pH 13), acetone, chloroform, cycioexane, ethanol, methanol and xylene. Reactivity to 30 % (v/v). H O was tested The insoluble pigment was examined ~y clu'omatographic analyus of the products obtained by permanganate oxidation and alkali fusion (Nieolaus et al., 1964). Cateehol production by P. rairabilis strain 297V was tested in the ethyl acetate extract of acidified growth supematant fluid (Shivprasad and Page, 1989). The free-radical nature of the pigment was investigated by the use of electron spin resonance (ESR) spectrometry measurements, carded out on a "Bruker ER 200 D" instrument equipped with a variable temperature accessory. Its possible antiradical properties were tested by estimating the reduction of the intensity of the superoxide anion ESR signal, using the procedures previously described (Sichel et al., 1981, 1991). The ultrastruetural features of the pigment were determined by scanning electron microscopy (SEM). A droplet of pigment suspension was layered on a cover glass, dried at 37°C and mounted on a stub using silver glue. A 5-nm gold layer coat was applied by an "Emscope SM300" (Emscope Laboratories, Ashford, Kent). Observations were carried out by an "S-4,000" field emission scanning electron microscope (Hitachi).

MATERIALS AND METHODS Bacterial strains

A wild-type P. mirabilis strain 297V, isolated from a human urine specimen, was used in this study. Escherichia coli strain JM10I carrying the pMC3A plasmid (Fuqua et al., 1991) was a gift from W. Claibome Fuqua, Comell University, Wing Hall Ithaca, NY 14853-8101. The laboratory strain E. coli Cla (Sasaki and Bertani, 1965) was used as a negative control in DNA analysis experiments. All strains were maintained in brain heart infusion medium (BHI) at-80°C.

2

Growth studies

Cultures were grown in Lufia-Bertani (LB) broth and LB broth containing 0.1% or 0,8% L-tyrosine and 250 mg/ml ferric pyrophosphate (Muller, 1985) to enable pigment production. A total volume of 50 ml of each medium was inoculated with 105 bacterial cells and incubated at 37°C in a rotary shaker for 48 h. Serially diluted samples, taken at intervals in duplicate, were plated out and incubated at 37°C for 24 h to determine the number of colony-forming units per ml. Production of the pigment was visually determined in the liquid growth medium and correlated with the stage of bacterial growth. Pigment isolation P. mirabilis 297V cultures were grown in LB broth containing 0.8 % L-tyrosine and 250 mg/ml ferric pyrophosphate at 37°C for at least 24 h, to allow maximal pigment production. Cells were removed by centrifugation at 2,500 g for 15 min and the supernatant was dialysed (2,000 molecular weight exclusion) against decreasing concentrations of NaCI (5 mg/ml; 2.5 mg/ml; 1 mg/ml) and dis-

ESR LB

= electronspinresonance. = Lufia-Bortani(medium).

"

°

.

The pigment's capacity to directly reduce ammoniac solutions of silver nitrate was tested by transmission electron microscopy (TEM). P. mirabilis 297V strain was grown on LB agar containing 0.1% and 0.8% L-tyrosine and 250 mg/ml ferric pyrephosphate for 48 h. Colonies were fixed in 3 % (v/v) glutaraldehyde in 100 mM phosphate buffer, sub-

I

SEM TEM

= =

scanningelectronmicroscopy. transmissionelectronmicroscopy.

STUDY OF A MELANIC PIGMENT OF PROTEUS MIRABILIS

jected to a modified Fontana-Masson assay (Sichei, 1988) and embedded using standard methods. Liver samples of Xenopus laevis were included for comparison. Observations were carried out on an "H7,000" transmission electron microscope (Hitachi).

DNA hybridization The presence of nucleotide sequences related to the melanin-biosynthesis-encoding operon (mel) from the marine bacterium S. colwelliana (Fuqua et aL, 1991) was examined by Southern hybridizatR,n (Southern, 1975). Whole-ceU DNA from P. mirabil/s strain 297V and E. coil Cla was obtained by bacterial lysis and purified by phenol-chloroform extraction (Sambrook et aL, 1989), After digestion with EcoRl and Pstl, restriction fragments were separated by eleetrophoresis in 0.8 % agarose horizontal slab gels and transferred to nylon membranes (ZetaProbe, Bio-Rad, Richmond, CA), with a vacuum blotter (Model 785, Bio-Rad, Richmond, CA). Nucleotide sequences related to the reel operon were detected by the use of a probe, consisting of the 1.3kb EcoRI-DraIII fragment of plasmid pMC3A (Fuqua et al., 1991). Plasmid pMC3A DNA from E. coli JMI01 was purified by phenol-chloroform extraction (Sambrook et al., 1989). The DNA restriction fragments used to generate the probe were separated by electrophoresis in horizontal slab gels containing 1% low melting point agarose. The relevant DNA restriction fragment~ from a slice of the gel, was labelled in the agarose by random primed incorporation of digoxigenin-dUTP (Boehringer-Mannheim, Germany)

according to the conditions recommended by the manufacturer. Hybridization and immunological detection procedures were carded out as previously described (Agodi et al,, 1990).

Inhibition of pigment production Growth features and pigment production of P. mirabilis 297V were also tested in the presence of

1 mM tropolone (Sigma Chemical Co, St Louis, MO), known as an efficient inhibitor of the enzyme tyrosinase (Kahn and Andrawis, 1985).

169

RESULTS Growth in complex media The amount of pigment produced increasea with the medium concentration of L-tyrosine. When LB containing 0.8 % L-tyrosine and 250 mg/ml ferric pyrophosphate broth cultures reached the logarithmic phase of growth, a coloured pigment, ranging from reddish brown to dark brown was t ~oduced. The pigment was detected after 4 h, and its intensity gradually increased, reaching peak levels after 8 h. Paral!e! cultures of the same strain in LB broth without supplements did not produce the pigment.

Characterization of the pigment Data obtained from the elemental analysis of P. mirabilis pigment and from the control sample of liver melanin from R, esculenta L. are shown in

table I. The resulting sample composition was in general agreement with the requirements of a melanin (Prom, 1992). Furthermore, the relatively low sulphur content and the relatively high nitrogen content suggested that the pigment could not be classified as a phaeomelanin or a catecholmelanin. The pigment retained in the dialysis tubing (2,000 molecular weight exclusion) was precipitated in 37 % HCI. It was soluble in alkaline water (pH 13), but only slightly soluble in either ethanol or methanol and insoluble in acetone, chlorophorm, cycloexane and xylene. It was bleached by the action of H202. Experiments on chemical degradation were performed in order to clarify the nature of the pigment. Given the limitations due to the weak sensitivity of the available methods (Nicolaus et

Table L Elemental analysis of the purified pigments from P. mirabilis 297V and Rana esculenta L. Iiver. Pigment

%C

%H

%N

%O

%S

% Fe

P. mirabilis Rana esculenta L.

60.51 48.98

4.72 3.10

6.05 8.99

27.90 37.78

0.32 0.25

0.24 O.10

170

A. AGODI ET AL.

al., 1964), only a positive result would have

been conclusive. Analysis of the ~2egradation products after permanganate oxidation of the pigment did not show the presence of any pyrrolic acid, nor did alkali fusion produce any catechoI, 3,5-dihydroxybenzoic acid or 2hydroxybenzoic acid. Moreover, no catechol production was found in the ethyl acetate extract of the acidified growth supematant fluid of P. rairabilis cultures. Thus, it was not possible to classify the pigment as an indolic melanin or as a catechol-melanln, nor was it possible to exclude this. Results from ESR studies are reported in figure 1. Based on figure I a, it is possible to state the stable free-radical nature of paramagnetic centres responsible for the ESR signal of the bacterial pigment. Moreover, the ability of the pigment to in vitro scavenge the superoxide anion is evident from the considerable reduction in the intensity of its ESR signal (fig. lb and lc). Thus, the physico-chemical characteristics of the pigment are those typically described for melanins (Sichel et al., 1981, 1991). Electron scanning micrographs showed the pigment to be organized as rounded aggregates of spherical bodies (fig. 2). Electron transmission micrographs of bacterial colonies subjected to a modified Fontana-Masson assay showed silver granules specifically precipitated inside the bacterial cells (fig. 3a and 3b). Analogously, eukaryotic pigmented cells, such as the Kupffer ceils of X. laevis tested, showed the silver granules specifically precipitated onto the melanosomes, the organelles responsible for melanogenesis (fig. 3c). Furthermore, from direct comparison of figures 3a and 3b, a relationship between the concentration of L-tyrosine, and the reducing activity of the pigment-producing bacteria can be seen. In conclusion, high resistance to hydrochloric acid as well as to various organic solvents, and reactivity to oxidizing agents, were differential (though not conclusive) chemical characteristics which, together with the positive response to the Fontana-Masson assay,

C

3000

3260

3600

3750

H (Q) Fig. 1. ESR spectra of (a) P. mirabilis 297V pigment, (b) supetoxide anion (pH 12 and 150 K) before the addition of P. mirabilis 297V pigment, and (e) superoxide

anion (pH 12 and 150 K) after 5-rain incubation with P. mirabilis 297V pigment.

correlated with the melanic nature of the pigment produced by P mirabilis from tyrosine.

DNA hybridization studies No nucleotide sequence related to the mel operon from the marine bacterium S. colwelliann was detected by Southern hybridization of the genomic DNA of P. rairabilis straili 297V.

S T U D Y O F A M E L A N I C P I G M E N T O F PROTEUS MIRABILIS

171

Fig. 2. Scanning electron micrograph of 13. mirabilis 297V pigment particular organized structure as rounded aggregates of spherical bodies (a), and at a higher magnification (b).

Inhibition of pigment production The effect of tropolone, ~mown as an efficient inhibitor of tyrosinase (Kahn and Andrawis, 1985), was determined to verify whether this enzyme, responsible for melanin production in many eukaryotic cells, is involved in the P. mirabilis pigment biosynthetic pathway. Growih and pigment production of P. mirabilis 297V were tested in the presence as well as in the absence of 1 mM tropolone. L-' LB medium c-Jntaining 0.8%

L-tyrosine, a brown diffusible pigmentation, the colour intensity of which increased after addition of ferric salts, was observed. L-tyrosine consumption was estimated visually from the diminution of the amount of the white amino acid deposit after 24-h incubation at 37°C. No pigment production was shown by P. mirabilis 297V 24-h cultures in LB medium containing 0.8% L-tyrosine and 1 nLM tropolone. Addition of ferric salts brought up only a reddish colour. Furthermore, no decrease in the white L-tyrosine deposit was detected.

172

A. AGODI ET AL. DISCUSSION

A satisfactory chemical definition of what constitutes a melanin is not available at present. As such, we can only suggest that, given its solubility features, bleaching by oxidizing agents and positive response to the Fontana-Masson assay, the pigment produced by P. mirabilis from tyrosine is melanic in nature. Furthermore, it shows the physico-chemical characteristic already described for other melanins: that of acting as a free radical trap, as shown by ESR measurements, For the first lime in the present study, as previously reported for synthetic melanins and melanins isolated from animal tissues, vegetable seeds and fungal spores (Siehel et al., 1981, 1991), this phenomenon was demonstrated for a bacterial melanin.

P o

" " ~ '::k.'i:I "... ,...,, • .

.

"

:

~i:~:~'7 :, ~

,

.

,

~

"

: ~ 2 : : :' ' : ~ - i ' : ~ : . .

,~" •

Attempts to chemically characterize the bacterial pigment failed to reveal whether it is an indolic, cateeholie or some other type of melanin (for a review see Prota, 1992).

:~::i,

SEM studies showed a particular organized structure of the pigment as rounded aggregates of spherical bodies• We hypothesize that such an unforeseen ultrastructure derives from an assembly of smaller into larger globes, but, unfortunately, our data are not sufficient to enable any hypothesis concerning the mechanism of pigment formation in vivo.

°..,

Fig. 3. TEM of P. mirabilis 297V colonies and mela-

nosomes. a) Colonies grown on LB a/;ar containing 0.1% L-tyrosine and 250 mg/ml ferric pyrophosphate. FontanaMasson assay. b) Colonies grown on LB agar with 0,8% L-tyrosine and 250 mg/ml ferric pyrophosphate. Fontana-Masson assay. c) Melanosomes contained in a Kupffer cell of X. laevis. Fontana-Masson assay.

To date, only two bacterial melanogenesis systems have been defined at the molecular level: the tyrosinase-encodinggenes from Streptomyces sp. (Bernan et ai., 1985; Huber et al., 1985; Muller et al., 1988) and the melanin-biosynthesisencoding genes from the marine bacterium S. colwelliana (Fuqua et al., 1991; Fuqua and Weiner, 1993), which have been cloned and sequenced. In the latterorganism, the me/A gene is responsible for melanogenesis via p-hydroxyphenylpyruvate hydroxylase (HPPH) enzymatic activityupon homogentisic acid (HGA) (Coon et al., 1994). Analogously, melanin production from tyrosine via the H G A pathway has been demonstrated in Pseudomonas aeruginosa (Yabuuchi and Ohyama, 1972),in V.cholerae and in a Hyphomonas strain(Kotob et al., 1995). D N A hybridization data did not reveal any nucleotidesequence

S T U D Y OF A M E L A N I C P I G M E N T OF PROTEUS MIRABILIS

related to the reel operon from the marine bacterium S. colwelliana in the P. mirabilis genome. This negative result let us e x c l u d e a c o m m o n mechanism of melanogenesis mediated by HPPH activity, On the contrary, we hypothesize a possible role of tyrosinase in P. mirabilis melanogenesis. Results obtained from experiments on pigment production inhibition by tropolone strongly support the involvement of this enzyme and possible role o f tyrosinase in P. mirabilis, Thus, as already demonstrated for other bacteria, the enzymarie system involved could be analogous to that found in eukaryotic cells. From this point of view, melanin formation may be regarded as a general pathway of evolutionarily distant organisms, Further experiments arc necessary to define the biological significance of the antiradical function in vivo, i.e. its possible role under oxidative stress conditions, such as those generated during phlogosis, or its relevance as a virulence factor. Moreover, D N A studies are in progress to define the molecular basis o f pigment production.

Acknowledgements We are grateial to Raffaele Bonomo, Michele Putfolio and Sebasfiano Sciuto for interesting suggestions and discussions. We are also indebted to W. Claibome Fuqua for providing £. coil strain JMI01 carrying pMC3A plasmid and to Ronald M. Weiner for encouragement and advice. This work was partially supported by grants from Ministero dell'Universita" e della Ricerca Scientifica (art. 65, DPR 382/80) and from Consiglio Nazionale delte Ricerche (n. 93.04668.CT04).

l~tude d'un pigment mtlanique de Proteus mirabilis Le pigment produit, en atrobiose, par Proteus mirabilis h partir des formes L de diffdrents acides amints aromatiques est-il de nature m~lanique ? I1 s'agit d ' u n pigment brun fonct, ayant plusieurs caract~res de la mtlanine: caract~res de solubilitt, d t c o l o r a t i o n par l ' a e t i o n d ' a g e n t s oxydants et positivit6 ~t la r~action de Fontana-Masson. Pour la p r e m i e r e fois, nous avons d t m o n t r 6 q u ' u n e mtlanine bacttrienne pout agir comme une trappe pour des radieaux libres, c o m m e cela a dtj/~ 6t6 montr¢ pour des mtlanines d'origines difftrentes. Des 6tudes en microscopic 61ectronique ~t balayage

173

montrent unc structure particuli~re du pigment de P. m i r a b i l i s organisEe en agrtgats arrondis de corps sphdriques. Les donndes d'hybridation ADNADN n'ont pus mis en 6vidence, darts le gtnome bacttrien, de stquence nucltotidique en correlation avec melA de St,ewanella colwelliana, un des deux syst~mes de mtlanogen~se dtjh dtfinis au niveau m o l t c u l a i r e chez les b a c t t r i e s . Des essais d'inhibition de la production de pigment suggtrent un r t l e possible de la t y r o s i n a s e clans la mtlanogen~se chez P. mirabilis. En conclusion, de l'ensemble de nos r~sultats il ressort que le pigment produit par P. mirabilis est de nature mtlanique. Mots-clis : Pigment, M61anine, Proteus mirabilis, M61anogen~se, Tyrosine; Electron spin resonance, Ultrastmcture, Trappe pour radicaux libres, Tyrosinase.

References Agodi, A., Jones, C., Threffall, E.J., D'Angelo, M. & Marranzano, M. (1990), Molecular characterization of trimethoprim resist0.'lce in Shigella sonnei in Sicily. Epidem, Infect,, 105, 29-40, Beman, V., Filpula, D., Herber, W., Bibb, M. & Katz, E. (1985), The nucleotide sequence of the tyrosinase gone from Streptomyces antibiotieus and characterization of the gone product. Gone, 37, 101-110. Coon, S.L., Kotob, S., Jarvis, B.B., Wang, S., Fuqua, W.C. & Weiner, R.M. (1994), Homogcntisic acid is the product of melA which mediates melanogenesis in the marine bacterium Shewanella colwelliana D. AppL Environ. Microbiol., 60, 3006-3010. Fuqua, W.C., Coyne, V.E., Stein, D.C., Chun-Mean, L. & Weiner, R.M. (1991), Characterization of melA : a gone encoding melanin biosynthesis from the marine bactelium Shewanella colwelliana. Gone, 109, I31136. Fuqua, W.C. & Weiner, R.M. (1993), The melA gene is essential for melanin biosynthesis in the marine bacterium ShewaneUa colwelliana. J. Gen. MicrobioL, 139, 1105-1114. Huber, M., Hinterrnann, G. & Lerch, K. (1985), Primary structure of tyrosinase from Streptorayces glaucescens. Biochemistry, 24, 6038-6044. Ivins, B.E. & Holmes, R.K. (1981), Factors affecting phaeomelanin production by a melanin producing (reef) mutant of Vibrio cholerae. Infect. Immun., 34, 895-899. Kahn, V. & Andrawis, A. (I985), Inhibition of mushroom tyrosinase by tropolone. Phytochemistry, 24, 905908. Katz, E., Thompson. C.J. & Hopwood, D.A. (1983), Cloning and expression of the tyrosinase gene from Streptomyces antibioticus in Streptomyces lividans. J. Gen. Microbiol., 129, 2703-2714. Kotob, S.I., Coon, S.L., Quintero, EJ. & Weiner, R.M. (1995), Homogentisic acid is the primary precursor

174

A. A G O D I E T AL.

of melanin synthesis in Vibrio cholerae, a Hyphomonas strain, and Shewanella colwelliana. AppL Environ. MicrobioL, 61, ! 620-1622. Lerch, K. & Ettlinger, L. (1972), Purification and characterization of a tyrosinase from Streptomyces glaucescens. Eur. £ Biochem., 31,427..437. Muller, G., Ruppert, S., Schmid, E. & Schutz, G. (1988), Functional analysis of alternatively spliced tyrosinase gene transcripts. EMBO J., 9, 2723-2730. Muller, H.E. (1985), Production of a brownish pigment by bacteria of the Morganella-Proteus-Providencia group. Zentralbl. BakterioL MicrobioL Hyg. Abt., 260, 428-435. Nicolaus, R.A., Piattelli, M. & Fattomsso, E. (1964), The structure of melanins and melanogenesis-lV. Tetrahedron, 20, 1163-1172. Pomerantz, S.H. & Murthy, V.V. (1974), Purifications and properties of tyrosinases from Vibrio tyrosinaticus. Arch. Biochem. Biophys., 160, 73-82 Prota, G. (1992). Melanins and malanogenesis. Academic Press Inc., London. Sambrook, J., Fdtsch, E.F. & Maniatis, T. (1989). Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory Press. Sasaki, I. & Bertani, G. (I965), Growth abnormalities in

Hfr derivatives of Escherichia coli strain C. J. Gen. Microbiol., 40, 365-376. Shivprasad, S. & Page, WJ. (1989), Catechoi formation and metanization by Na+-dependent Azotobacter chroococcum: a protective mechanism for aeroadaptation? Appl. Environ. Microbiol., 55, 1811-1817. Sichel, G. (I988), Biosynthesis and function of melanins in hepatic pigmentary system. Pigra. Celt Res., 1,250-258. Sichel, G., Brai, M., Palminteri, M.C. & Sciuto. S. (1981), Seasonal dependence of ESR features of frog melanins. Comp. Biochem. Physiol., 7013, 611-613. Sichel, G., Corsaro, C., Scalia, M., Di Bilio, A.J. & Bonomo, R.P. (1991), In vitro scavenger activity of some flavonoids and melanins against O~. Free Radic. BioL & Med., 11, I-8. Southern, E.M. (1975), Detection of speeific sequences among DNA fragments separated by gel electrophoresis. J. Mol. Biol., 98, 503-517. Yabuuchi, E. & Ohyama, A. (I972), Characterization of "pyo-melanin"-produeing strains of Pseudomonas aeruginosa. Int. J. Syst. Bacteriol., 22, 53-64. Yoshimoto, T., Yamamoto, K. & Tsuru, D. (1985), Extracellular tyrosinase from Streptomyces sp. KY-453: purification and some enzymatic properties. J. Biochem., 97, 174%1754.